There’s been very little attention in Australia given so far to a new study on the cost of the Iraq war, summarised in this op-ed piece in the Los Angeles Times:

Economists Linda Bilmes and Joseph Stiglitz estimate the total cost at
more $US1 trillion, and possibly as much as US$2
trillion if the occupation lasts until 2010 – or something like 20
times the figure the US government predicted beforehand.

There is a twofold lesson in this. First, the monumental dishonesty
associated with the Iraq war. Even assuming that its goals were genuine
and achievable, and that they were worth paying a high price for, there
is no evidence that alternative ways of pursuing them (at a fraction of
the cost) were seriously considered.

But there’s also a more general lesson about how government works.
Bilmes and Stiglitz say that “Elsewhere in the government, it is
standard practice to engage in an elaborate cost-benefit analysis for
major projects.” It’s certainly more common than it once was, but even
now most such analyses are not worth the paper they’re printed on. In
nine out of ten major projects, the political decision comes first, and
the figures are then found (or invented) to justify it.

Especially common in government is the failure to consider longer-term
costs. Bilmes and Stiglitz list some of them: “lifetime healthcare and
disability benefits for returning veterans … increased cost of
replacing military hardware … large reenlistment bonuses … rising
interest cost on the extra debt … economic costs of injury and death
… macro-economic costs such as the effect of higher oil prices.”

I remember from my own time in government the difficulty in getting
ministers and bureaucrats to understand how pay rises, for example, had
a long-term aspect: that there was not just an initial cost to the
budget, but also an increased superannuation liability that would only
become apparent years later. Compared to the Iraq war it was small
beer, but the fundamental problem was the same – political
decision-makers focus on the big bang and the short term.